1. Field
Example embodiments relate to a method for detecting defects in a substrate having a semiconductor device thereon. Other example embodiments relate to a method for detecting defects in a substrate having a semiconductor device thereon using a less complex process within a shorter amount of time.
2. Description of the Related Art
As processes for manufacturing a semiconductor device become more complex and critical dimensions of semiconductor devices decrease, problems may occur such as defects (e.g., bridges) between semiconductor patterns, breaking of the semiconductor patterns, voids in the semiconductor patterns, etc.
A process for detecting defects generated during processes for manufacturing a semiconductor device may be performed after performing each unit process thereof. If the detection process is accurately performed, successive processes may be performed without defects, preventing a decrease in a yield rate.
Comparison algorithms for detecting defects in a semiconductor device may be classified as a random mode algorithm or an array mode algorithm.
A random mode uses a die-to-die method in which a target die to be inspected is compared to an adjacent die. In the random mode, a first region of the target die may be compared to a second region of the adjacent die, which corresponds to the first region. If pixels in the first region are found to be different from those in the second region, defects may be present. In the random mode, corresponding regions between adjacent dies are compared with each other in order to detect defects in a region having irregular patterns therein and defects in a region having regular or repeated patterns therein. If defects are detected in the random mode, noise may increase due to a color difference or a focus difference between the adjacent dies. As the noise increases, detection power decreases.
In an array mode, images of repeated patterns on a die may be compared to one another at a predetermined interval. If a difference is found in the images of the repeated patterns, defects may be present. That is, a pixel having the different images may have defects. In the array mode, adjacent patterns on a die are compared to one another in order to detect defects in a region having the repeated patterns therein.
There are seldom color differences or focus differences between the adjacent patterns such that defects may be detected in the array mode more accurately than the random mode because adjacent regions, in which the repeated patterns are formed in one die, are compared to one another in the array mode. The signal-to-noise ratio (SNR) of the array mode may be about 1.5 times higher than that of the random mode. Minute defects may be accurately detected in the array mode due to the higher SNR thereof. Defects in a substrate having a highly-integrated semiconductor device thereon are usually detected using the array mode because the detection power of the array mode is higher than that of the random mode.
A semiconductor device may include a first region where patterns are repeatedly formed along an x-axis direction and a second region where patterns are repeatedly formed along a y-axis direction. The first and second regions, in which repetitive patterns are formed in different directions from each other, are inspected.
A substrate having a semiconductor device thereon is scanned in the x-axis direction with an image sensor so that images are obtained of first repetitive patterns in a first region, in which the first repetitive patterns are formed in the x-axis direction. The obtained images of the first repetitive patterns are compared to one another at predetermined intervals so that defects in the first region may be detected. After rotating the substrate by about 90 degrees in the same plane, the substrate is scanned in the y-axis direction with the image sensor so that images are obtained of second repetitive patterns in a second region, in which the second repetitive patterns are formed in the y-axis direction. The obtained images of the second repetitive patterns are compared to one another at predetermined intervals so that defects in the second region may be detected.
A time delay integration (TDI) sensor chip used in the array mode may be arranged in such a way that images of patterns may be obtained only in a scanning direction. As such, defects may be detected in a region where the patterns are repeatedly formed in the scanning direction. An inspection time may be longer if defects of the substrate are detected using the array mode, in which patterns are repeatedly formed on the substrate in the x-axis and y-axis directions, because the substrate is scanned in each of the x-axis and y-axis directions in order to detect the defects.
A peripheral portion of the region, in which the repetitive patterns are formed, may have a shape different from that of another portion of the region. As such, detecting defects in the peripheral portion of the region may be difficult using the array mode, in which images of the repetitive patterns in one portion of the region are compared to those of another portion of the region.